cancel
Showing results for 
Search instead for 
Did you mean: 

HPET High Precision Event Timer

lenehan
Level 10
Hi all,

Found this two post from different web site intersting worth reading. cheers.

http://forums.guru3d.com/showthread.php?t=368604

http://www.neowin.net/forum/topic/1075781-tweak-enable-hpet-in-bios-and-os-for-better-performance-an...
Change is the law of life. And those who look only to the past or present are certain to miss the future.
87,490 Views
20 REPLIES 20

jollywombat
Level 10
lenehan wrote:
Hi all,

Found this two post from different web site intersting worth reading. cheers.

http://forums.guru3d.com/showthread.php?t=368604

http://www.neowin.net/forum/topic/1075781-tweak-enable-hpet-in-bios-and-os-for-better-performance-an...



Interesting, I will have to go play with this and see how it affects my system. Alot of back and forth there on if its good or not it seems. Nice find! :cool:

m_oreilly
Level 7
much faster boot with timer disabled in bios ;). no issues so far...
Asus Rampage IV GENE
i7 3930k @ 4.8 24/7
custom loop
16 gigs Trident X
HIS HD7970
2x OCZ Vector 256gb raid0 via UEFI
Crossover 27Q
Win8 pro

jollywombat
Level 10
Tested it out with three test sets. Computer is Windows 8 RTM 8400 64bit, 3930k @ 4.5ghz, RIVE, 2x240gb Corsair GT Raid-0, 2x GTX680 SLI, 32gb 4x8gb DDR3 Corsair Dominator.

1) Disabled in bios - verified disabled in windows

2) Enabled in Bios - windows reporting it was not using it via WinTimer& bcdedit /enum

3) Enabled in Bios - verified enable with windows via bcdedit & WinTimer

All three averaged out to the same boot time of 17.9-18.1 seconds. Measured with Windows 8 taskmanager startup logging. Restarted the computer 10 times in each mode, recorded times, then averaged them.

Tested with superpi to see if it would make any different in time to calculate, no difference between any of them.

Verified stability with intel burn test on very high for 10 runs on each, but as expected, all were stable.

Tested 3DMark 11 in all three settings, 5 runs each setting. Averages for all were fairly similar, not exact but test results from 3dmark always vary +/- 100 score typically.

General windows experience felt the same, no noticeable difference between any settings.


Would be interested in anyone else who can do some good testing here to compare with.

m_oreilly
Level 7
unless you disable the hiber/fast boot of win8, you won't be able to notice anything. win8 is an awful excuse for an os as it stands
Asus Rampage IV GENE
i7 3930k @ 4.8 24/7
custom loop
16 gigs Trident X
HIS HD7970
2x OCZ Vector 256gb raid0 via UEFI
Crossover 27Q
Win8 pro

m.oreilly wrote:
unless you disable the hiber/fast boot of win8, you won't be able to notice anything. win8 is an awful excuse for an os as it stands


Hibernate is disabled on the computer, and the hyberfil.sys was removed. Did not mess with any fastboot options however.

I will be putting Windows 7 back onto the second partition here soon. Had formatted recently and had not gotten both OS'es back up yet for testing between the two. I know 3dM11 likes win7 far more than Win8, generally get 1000+ more physics score with same settings between the two.

Will test again once win7 is back on.

In the mean time, anyone with actual performance data care to chime in for what they saw? Would be useful for folks to see results

wow, even streaming directv2pc is smoother. winning here. i think if you notice improvement in your system response, there is some merit. i also think we would see differing results per rig. i am on the latest bios, 3.1 orom, no intel management engine, no C states, minor bios adjustments to bring system to 4.8 24/7, running off a raid0 ssd array, server 2008r2. i did notice when i disabled the timer, the board shut itself down, as if i had initiated some major clock change or the like.
Asus Rampage IV GENE
i7 3930k @ 4.8 24/7
custom loop
16 gigs Trident X
HIS HD7970
2x OCZ Vector 256gb raid0 via UEFI
Crossover 27Q
Win8 pro

m.oreilly wrote:
wow, even streaming directv2pc is smoother. winning here. i think if you notice improvement in your system response, there is some merit. i also think we would see differing results per rig. i am on the latest bios, 3.1 orom, no intel management engine, no C states, minor bios adjustments to bring system to 4.8 24/7, running off a raid0 ssd array, server 2008r2. i did notice when i disabled the timer, the board shut itself down, as if i had initiated some major clock change or the like.


Hi mate, it would be nice if you can share you congfig detail here. Thanks.
Change is the law of life. And those who look only to the past or present are certain to miss the future.

*note, getting some indications server 2012 has it on if available, unsure if 2008 then had it enabled through a patch or not, still think you have to specifically enable it though

jollywombat
Level 10
oreilly,

To the best of my understanding, even though this is enabled in the bios, it was not in use in widows versions 7,8, server 08, but vista had it enabled. So it should not have had an effect unless you had previously specifically enabled it as the only timer in the OS. Just curious if you had done that or checked it with Wintimer and such here so we can see what your results were for comparison here.